1. Removing the noose does not lead to increased THD. In the case it was mentioned that it did in this thread, the full modification had not been performed at that point. Once the modification was completed, distortion was no higher than before, and is usually less.
2. Lowering the voltage on the AF amplifier stage allows the inverter stage to bias itself properly within the linear portion of its operating curves.
3. No. See below.
The noose performs two functions: (1) It limits AC drive to the output stage, possibly as a protection element for the output tubes, and (2) It allows the inverter stage to bias itself below saturation levels in spite of the higher than normal plate voltage offered by the previous stage that is direct coupled into it. As such it was a move to maximize gain production in the AF Amplifier stage, although at the expense of greater than normal losses through the inverter stage due to the noose causing it to bias itself so cold.
The problem with the noose is therefore two fold: (1) For output tubes requiring an unusually high amount of grid bias voltage, it puts the maximum power drive requirements of the output stage outside of that which the stock inverter design is capable of, causing increased distortion at higher power output levels, and (2) Because the inverter section biases itself so cold, many Russian based 12AX7 tubes will not operate properly in the stock design since the quiescent operating point is so low on the load line -- a point at which characteristic consistency from one tube to another varies considerably. Russian 12AX7s are very good tubes when operated throughout the linear portion of the operating curve. But near the cut-off end of the curves, the Russian tube characteristics vary enough -- as compared to the original American pieces -- as to be problematic in the stock design.
By removing the noose then: (1): The inverter stage is able to produce plenty of low distortion drive to accommodate output tubes requiring even very high bias voltage levels, and (2), The stage will operate properly with American or Russian based tubes, because the inverter stage then operates well within the linear portion of its grid curves. The OLG (Open Loop Gain) of the overall design remains unchanged, as the loss of gain in the AF Amplifier stage due to operating it from a slightly lower B+ source is more than made up for by eliminating the excessive losses in the original inverter stage since it is now operating well within the linear portion of its curves. The result is that distortion -- when properly implemented -- is invariably reduced.
The modification then has nothing to do with trying to outwit the clever pioneering engineers of Fisher, but rather, adapt the circuit to be more in align with the reality of today's vacuum tube audio environment.
Dave